Epidemiology, Susceptibility, and Risk Factors Associated with Mortality in Carbapenem-Resistant Gram-Negative Bacterial Infections Among Abdominal Solid Organ Transplant Recipients: A Retrospective Cohort Study

All the electrical medical records of 1452 ASOT recipients from August 1, 2013, to August 1, 2020, were retrospectively reviewed. As Fig. S1 shows, a two-part analysis was performed (1) to evaluate the risk factors associated with 90-day mortality and (2) to identify the drug resistance of CR-GNB pathogens.

The retrospective cohort study was performed at the Third Xiangya Hospital of Central South University, an 1800-bed tertiary-care teaching hospital with a long-term ASOT program (annual average of 180 kidney and 35 liver transplantations), Changsha, China.

Patients < 16 or > 80 years old were excluded. There were only two sources of organs harvested for transplantation: living-related donors and donors after citizen’s death. The prophylactic regimen used 1 h before transplantation was beta-lactamase inhibitors/semisynthetic penicillin, second- or third-generation cephalosporins, or carbapenems. According to the result of the culture before transplantation, a precise antibiotic was given within 72 h after transplantation. The treatment we recorded was the initial regimen after obtaining the microbiology profile. Tigecycline was used at 50 mg every 12 h (100 mg the first time) and mostly in the abdominal cavity or deep wounds and skin infections instead of bloodstream infections as monotherapy. High-dose (2 g every 8 h), extended-infusion meropenem (over 3 h) was considered an appropriate therapy for initial treatment. Due to limited drug resistance tests of polymyxin and ceftazidime-avibactam, prescriptions of both antibiotics as effective options depended on the patient's condition. Data on the subsequent microbiology profile and corresponding treatment were not collected. For patients experiencing more than one episode of CR-GNB infection, only the first episode was included.

The onset of CR-GNB infections referred to the collection date of the first positive culture with clinical evidence of infections. Patients with positive microbiology results, according to the criteria of the Centers for Disease Control, were defined as having infections [8]. Donor-derived infections were defined as the same infections that were transmitted from the donor to more than one recipient [9].

Appropriate empirical antimicrobial therapy indicated that antibiotics susceptible to CR-GNB in vitro were prescribed to treat suspected CR-GNB infections within 48 h after having drawn the cultures of specimens. Treatment during the initial stage of infections was categorized as monotherapy or combination therapy. Combination therapy was defined as treatment with two or more agents active in vitro against the infecting strain. Septic shock was diagnosed in recipients with CR-GNB infections who required vasopressor therapy to maintain mean arterial pressure of at least 65 mmHg with serum lactate level > 2 mmol/l after adequate fluid resuscitation [10]. Crude mortality included all causes of death within 90 days after the onset of infections.

Identification and antimicrobial susceptibility testing of CR-GNB were carried out using the Vitek-2 system (bioMérieux, Marcy L’etoile, France). Antimicrobial susceptibility was determined by the Kerby-Bauer disk diffusion method, and the minimum inhibitory concentration was measured by agar dilution in the National Committee for Clinical Laboratory Standards guidelines. All the antimicrobial agents were products of Oxoid Ltd., Hampshire, UK. Intermediate susceptibility to the antibiotics was considered as resistance. The strain was considered to be CR-GNB when the minimal inhibitory concentration of meropenem or imipenem was ≥ 2 mg/l [11].

The clinical and demographic characteristics included: age, sex, hallmarks of infections (temperature and white blood cell count), site of infection, re-operation, mechanical ventilation, hospital-acquired infections, septic shock, empirical antimicrobial therapy, type of transplantation, and treatment regimen. Laboratory records, collected within the first 24 h after the culture was drawn, included: the serum albumin level, total bilirubin and creatinine levels, and platelet and lymphocyte counts.

The follow-up period for all the patients was 3 months (90 days) after the onset of infections. The patients with < 3 months of follow-up were excluded from the cohort. The clinical outcomes were divided into death and alive for risk analysis.

The Institutional Review Board of the Third Xiangya Hospital of Central South University endorsed this study protocol prior to data collection (number: 2020-S629). The Institutional Review Board approved the waiver of informed consent from patients because this was a retrospective cohort study where information was obtained from electrical medical records and this study did not directly interfere with the enrolled patients as the data were de-identified and anonymously analyzed. Our study was conducted according to the Declaration of Helsinki and the Declaration of Istanbul, and no donors influenced or paid for our study. As a privacy statement, the authors guarantee the confidentiality of the patient data.

For continuous variables, data were expressed as median (25–75% quartile) or mean (± SD). Categorical variables were compared with the χ² test or Fisher exact tests. Univariate analysis was applied to inspect the association between variables and mortality caused by CR-GNB infections. The follow-up time was analyzed and defined as days alive from the onset until day 90, and a Cox regression (forward likelihood ratio method) analysis was performed to model hazards using this time variable. Hazard ratio (HR) and 95% confidence intervals (CI) were calculated to assess the strength of all relations. The Kaplan-Meier curve was used to describe the survival distribution, and the log rank test was used to compare survival time among all the independent risk factors. Statistical significance was defined as P < 0.05 (two-tailed), and all analyses were carried out using SPSS 24.0 (IBM SPSS Statistics, IBM Corp., Armonk, NY, US).

During the 8-year cohort study period, CR-GNB infections occurred in 153 of 1452 (10.5%) ASOT recipients including 1192 kidney and 260 liver recipients in the Third Xiangya hospital. The kidney and liver transplantation prevalence of CR-GNB infections was 8.7% and 18.8%, respectively.

Table 1 shows the clinical characteristic, demographic, and laboratory data of patients with CR-GNB infections. The median age of 153 ASOT recipients was 43.3 years, with 141 grafts from donation after citizen's death and 12 grafts from living-related donors. Regarding the type of CR-GNB infections, urinary tract infection was the most common (31.4%), followed by bacteremia (19.6%) and pneumonia (19.6%). Of these 153 patients, 82.4% (126 of 153) suffered from nosocomial origin infections and 65.4% (100 of 153) were infected by XDR pathogens. One hundred eight patients (82.4%) did not receive appropriate empirical antibiotic treatments within 48 h after the onset of CR-GNB infections. Twenty-one (13.7%) patients required mechanical ventilation support as a treatment for respiratory failure caused by the infections, and 18 (11.8%) patients developed septic shock at the onset of infections. Twenty (13.1%) patients acquired infections from the donor, and 57 (37.3%) patients experienced multiple organism or site infections. The crude 90-day mortality rate was 23% (15 of 153).

The comparison of death and alive groups is shown in Table 2. Nosocomial infections (P = 0.035), multiple infected organisms or sites (P = 0.003), intensive care unit stays > 7 days (P < 0.001), mechanical ventilation (P < 0.001), pneumonia (P = 0.023), septic shock (P < 0.001), platelet count < 50,000/mm³ (P < 0.001), albumin < 30 g/l (P = 0.023), creatinine > 1.8 mg/dl (P = 0.042), total bilirubin > 2.5 mg/dl (P = 0.001), combination therapy (P < 0.001), and XDR (P = 0.018) were significantly different between the two groups in the univariate analysis. In the multivariate analysis, mechanical ventilation (HR 7.40, 95% CI 2.69–20.38, P < 0.001), septic shock (HR 7.41, 95% CI 2.86–19.23, P < 0.001), and platelet count < 50,000/mm³ (HR 4.00, 95% CI 1.49–10.76, P = 0.006) were three independent risk factors associated with crude mortality in Table 3. Figure 1 shows the survival time of patients with independent risk factors. For the outcome of patients with septic shock versus non-septic shock, the survival rate in the septic shock group was significantly lower (38.9% vs. 90%, P < 0.001). The survival rate was significantly lower in the patients requiring mechanical ventilation than in the patients without mechanical ventilation at the onset of infections (28.6% vs. 93.9%, P < 0.001). Compared with the platelet count ≥ 50,000/mm³ group, the survival rate was significantly lower in the platelet count < 50,000/mm³ group (50% vs. 91.5%, P < 0.001).

Figure 2 shows in detail the classification and percentage of the respective microorganisms. The most common pathogen of CR-GNB infections was CRAB (n = 47, 31%), followed by CRKP (n = 39, 25%), carbapenem-resistant Escherichia. coli (n = 20, 13%), CR-Flavobacterium (n = 16, 10%), and others. Table 4 shows the antibiotic resistance rates of CR-GNB via a bacterial category. The drug resistance rates of CRAB and CRKP pathogens to all antibiotics tested were both > 50% except for tigecycline (TGC). The CR-GNB pathogens were relatively susceptible to TGC with a drug resistance rate as low as 33.3% and highly resistant to other categories (> 60%). Figure 3 shows the comparison of resistance rates of CR-GNB to ten commonly used antibiotics during two halves of the period (2013–2016 vs. 2017–2020). A significantly increasing resistance trend of the TGC, from 24 to 40%, was observed in our cohort as time went on (P = 0.04).